q yao comments

Results 321 comments of


                                            q yao

Why official torch2trt doesn't invoke any plugin, but this invoke too many plugins even with simple model

Hi In these 'simple model' case, It is caused by `adaptiveavgpool2d`/`adaptivemaxpool2d`, Here is the difference | | official | onnx | this | |----| ----| ----| ----| |static| o |...

Why official torch2trt doesn't invoke any plugin, but this invoke too many plugins even with simple model

@ideafold without plugin

some issues when build with win10 use clion cuda11.1 tensorrt 8.0

Thanks for the notification! Would you like to create a PR to fix it?

Hello. Thx for your works. I have problem with build cmake

Did you download the TensorRT from official site?

hi, error: identifier "nullptr" is undefined

Hi, My gcc version is 7.5.0, try update gcc.

hi, error: identifier "nullptr" is undefined

@svmsunny thanks for the advice. I will update cmake version later.

hi, error: identifier "nullptr" is undefined

@Maxfashko Thanks for the advice. I am not so familiar with linux kernel. Did that means there will be an ABI error when compile the project with gcc-9 ? Will...

how to add a converter for custom operateion written by c++

- First, you need to implement the custom ops you need as a TensorRT plugin, [here](https://github.com/grimoire/amirstan_plugin/tree/master/src/plugin) are some examples. - After you have your plugin, add a creator in [torch2trt_dynamic/plugins](https://github.com/grimoire/torch2trt_dynamic/tree/master/torch2trt_dynamic/plugins),...

fix interpolate dynamic input

Hi Thanks for the bug report. The dynamic shape support of interpolate is a little bit ... complex. I warp the Tensor.shape[i] to a IntWarper in [size.py](https://github.com/grimoire/torch2trt_dynamic/blob/master/torch2trt/converters/size.py) to trace the...

Error number of channels in input tensor to a convolution layer must not be dynamic

Hi, according to the error log you provided, is there some structure like `torch.nn.conv2d(input, weight,...)` in your model where weight is not a static parameter?